|
From: | Judd Storrs |
Subject: | Re: Same .m file: different results with different versions of Octave |
Date: | Mon, 19 Apr 2010 11:44:35 -0400 |
I'm not sure that's really a good fix for the problem. GSL's approach is fundamentally superior because it avoids cancellation entirely by factoring the results into a product. All of the additions in GSL's implementation are between probably positive values that cannot have cancellation problems. glibc's algorithim is comparatively naive and should just be replaced but I haven't had a chance to verify the math behind the GSL terms yet. --judd
[Prev in Thread] | Current Thread | [Next in Thread] |